[04:19] <carlos> "morning"
[04:19] <daf> heh :)
[04:19] <daf> did you have an exam?
[04:20] <carlos> daf: I had, but I was not able to do it
[04:20] <carlos> I did not finished a program I need to do the exam
[04:20] <carlos> and I went to sleep at 6:00AM
[04:20] <daf> ouch
[04:20] <carlos> yes, really sad
[04:20] <carlos> I hate powerbuilder!!!
[04:20] <carlos> it's the second time I have this problem
[04:20] <kiko> there are worse things in life
[04:21] <carlos> kiko: I know
[04:21] <daf> powerbuilder?
[04:21] <carlos> daf: yes, the program they request use should be developed with powerbuilder
[04:21] <carlos> grr, I really hate xchat's autocompletation
[04:22] <carlos> or perhaps was my problem ... 
[04:22] <carlos> well, back to work, that's the important thing...
[04:23] <daf> indeed
[04:23] <kiko> SteveA, debonzi is hung up with tla again
[04:24] <kiko> SteveA, any hope we can get some traction on the issue asap?
[04:24] <daf> carlos: shall we talk about tasks for a bit?
[04:24] <carlos> daf: yes, please. 
[04:24] <daf> kiko: I think it's dependent on ddaa
[04:24] <kiko> daf, ugh, really?
[04:24] <daf> kiko: from listening in #canonical
[04:24] <carlos> daf: We should add all remaining ones to bugzilla so we could see what's missing easily
[04:25] <daf> carlos: good idea
[04:25] <carlos> daf: here or should we move to #rosetta?
[04:25] <daf> I think here is ok
[04:26] <carlos> ok
[04:27] <daf> what are we missing from Bugzilla?
[04:28] <SteveA> kiko: I saw lifeless and ddaa discussing it on #canonical.  I saw that lifeless asked ddaa to make it a priority.
[04:29] <kiko> okay, because we're blocked by tla (once again)
[04:29] <SteveA> carlos: we're not using #rosetta any more
[04:29] <carlos> daf: the import of all projects we have already inside arch
[04:29] <carlos> SteveA: ok
[04:30] <SteveA> kiko: you can still write code, although it will be outside the SCM system.
[04:30] <daf> carlos: hmm, this doesn't really feel like something that belongs in bugzilla
[04:30] <carlos> it's a task
[04:30] <kiko> SteveA, writing code and not committing frequently with arch seems to be a recipe for disaster in our experience :)
[04:30] <carlos> if we are going to use it as task/bug tracking, I think it belongs to bugzilla
[04:31] <SteveA> kiko: I can't offer any help above what ddaa can do for you.
[04:31] <daf> I suppose so, but it feels like abusing bugzilla a bit :)
[04:31] <kiko> sure, I'm just communicating our situation.
[04:31] <carlos> daf: It's something that blocks the alpha release :-)
[04:31] <SteveA> thanks.  I understand that it is extremely frustrating.
[04:31] <daf> carlos: is there anything else?
[04:31] <carlos> daf: we also need to add "Request new project page"
[04:32] <kiko> we'll find some more offline work we can do meanwhile.
[04:32] <SteveA> ok, thanks.
[04:32] <carlos> and " View recently translated projects on the translator dashboard."
[04:32] <daf> okay, can you make it a task to file bugs for these things?
[04:33] <carlos> daf: and I think that nothing more is missing from RosettaAlpha page
[04:33] <carlos> daf: sure.
[04:33] <carlos> well, the import projects from arch...
[04:33] <carlos> has some tasks like the scripts to do it
[04:34] <carlos> the main one that imports .po/.pot files is done 
[04:34] <carlos> but we need one to create projects and products
[04:34] <daf> I need to file a bug for the import code
[04:34] <daf> the import script isn't working for me: all the sequence numbers in the template I import are 0
[04:35] <daf> I think it's some sort of SQLObject problem
[04:35] <carlos> daf: I filed some bugs about the export also
[04:35] <carlos> seems like the code is not in sync with latest interface
[04:35] <daf> I think I've fixed the export
[04:35] <daf> but the traversal is not working
[04:36] <carlos> did you saw bugzilla's bug #1912? is it fixed ?
[04:36] <daf> yeah, I saw it
[04:36] <carlos> daf: I was not able to test the traversal because the export bugs
[04:36] <daf> I think it is fixed
[04:36] <daf> I will commit and merge soon
[04:37] <carlos> daf: ok, I will review then the traversal code and finish it
[04:37] <carlos> is the #1909 also fixed?
[04:37] <carlos> (the one about the % problem)
[04:37] <SteveA> back in about an hour
[04:37] <daf> it hasn't been merged yet
[04:38] <daf> stub has a fix which we can use
[04:38] <daf> if we use his Zope branch, there's no problem
[04:39] <carlos> daf: then I think that's the problem with the traversal to export .po files. But I will review it in case it's another problem...
[04:40] <carlos> daf: what's lalo working on now?
[04:40] <daf> I don't know
[04:40] <daf> I haven't heard from lalo since Friday
[04:40] <carlos> Could I assume that any task that it's not assigned at bugzilla to lalo is free to implement?
[04:40] <carlos> hmm
[04:40] <carlos> ok
[04:41] <daf> at the moment, anything is up for grabs
[04:41] <daf> feel free to assign bugs to yourself
[04:41] <carlos> daf: who is importing code to arch?
[04:41] <carlos> because I suppose we need to talk with them and know how to get the source to feed our database
[04:42] <daf> not sure: lifeless or lamont or thom?
[04:42] <carlos> daf: I forgot another task, we need to create the webpage with the list of projects/products from Ubuntu 
[04:43] <daf> okay, it's another bug then
[04:43] <carlos> well, and of course, we need "limi's magic" to all pages before release
[04:44] <carlos> I will see if he has a bugzilla account
[04:44] <carlos> and I will add also that bug
[04:45] <daf> the export fix looks like this: http://muse.19inch.net/~daf/misc/export-fix.patch
[04:46] <daf> does it look sane?
[04:48] <daf> carlos: ?
[04:48] <carlos> daf: looking
[04:49] <carlos> daf: yes, I think it's valid
[04:49] <daf> in that case, I'll merge it now
[04:50] <carlos> my question about it is, should that file use the sqlobject api or only the interfaces?
[04:50] <daf> the exporter should not know anything about SQLObject, I think
[04:50] <daf> it shouldn't need to
[04:51] <carlos> ok
[04:51] <carlos> daf: anything else?
[04:52] <carlos> daf: who should be the initial owner of the tasks I'm going to add?
[04:53] <daf> me
[04:53] <carlos> ok
[04:54] <daf> unless you know it's something you're going to work on yourself
[04:54] <daf> in which case it's you :)
[04:54] <carlos> :-P
[04:54] <carlos> daf: we should use the QA field
[04:54] <daf> how does that work?
[04:55] <carlos> daf: it's someone that gets the bug reports like the owner and it's useful to track the fixes
[04:55] <carlos> and follow the status of the tasks
[04:56] <daf> it's like a CC?
[04:56] <carlos> usually,  the project owner is the one added there
[04:56] <carlos> daf: more or less
[04:56] <carlos> daf: but it's added by default always
[04:57] <daf> ok
[04:57] <daf> where do we set it?
[04:57] <carlos> I don't know if you have the rights to do it, you should edit the launchpad product
[04:57] <carlos> and select the rosetta module
[04:58] <daf> I'm not sure if I have rights to do that
[04:58] <carlos> there you have a default assigned and the default QA
[04:58] <carlos> another option is to use an alias for the team so we can know the status of the bugs
[04:59] <daf> let's try the QA thing for now
[04:59] <daf> justdave: about?
[04:59] <justdave> daf: yep
[05:00] <daf> justdave: did you follow that about QA?
[05:01] <justdave> not exactly...  you want someone added to the qa for rosetta, but I didn't follow who. :)
[05:01] <justdave> conversation looks like you, but you're already the owner
[05:01] <justdave> but that would make sense if you're trying to stay in the loop even if someone else takes the bug
[05:02] <kiko> or maybe a rosetta-bugs mailing list so everyone can stay in the loop if they want to.
[05:04] <daf> kiko: hmm, maybe
[05:04] <daf> https://bugzilla.warthogs.hbd.com/bugzilla/chart.cgi?category=-All-&subcategory=-All-&name=181&select0=1&label0=Open+Launchpad+bugs&line0=133&datefrom=&dateto=&action-wrap=Chart+This+List
[05:06] <daf> justdave: is it possible to save charts?
[05:06] <daf> justdave: it would be nice to have "open Launchpad bugs" and "open Rosetta bugs" charts
[05:07] <justdave> people keep asking that, and Gerv keeps telling them "just bookmark it"
[05:07] <justdave> I had to fight him to keep saved queries :)
[05:07] <daf> saved queries are very useful
[05:08] <daf> I think application-specific bookmarks make sense
[05:08] <daf> it's the same in Rosetta - you want to find certain things quickly
[05:08] <justdave> Easiest thing for us to do right now is probably just collect the URLs for the ones we want, and stick up a static page with the links on it.
[05:09] <daf> that would do the job
[05:22] <carlos> daf: I think we have all tasks now at bugzilla
[05:23] <daf> I hope so :)
[05:24] <carlos> daf: :-P
[06:10] <lalo> hello
[06:12] <jblack> debonzi: ping
[06:15] <kiko-fud> jblack, he's out of office right now, how's it going?
[06:16] <jblack> Its going well. I've got a fix ready for him.
[06:16] <lalo> hey James
[06:16] <jblack> lalo! 
[06:16] <kiko-fud> jblack, cool, can you use email, perhaps to launchpad, summarizing and providing the fix?
[06:17] <jblack> Um, well, the thing about this sort of fix is that its probably not right for you guys to do it unattended.
[06:17] <jblack> the solution involves artfully dodging race conditions.
[06:18] <jblack> For a longterm fix, tla needs to change.
[06:18] <kiko-fud> maybe I don't want to know after all :)
[06:18] <spiv> Yow :)
[06:22] <SteveA> oh.  we broke tom lord's arch...
[06:22] <jblack> Nope.
[06:22] <carlos> lalo: hi
[06:22] <lalo> hey carlos
[06:23] <lulu> lalo: howzit
[06:23] <jblack> You guys are just frequently bumping into a edge case in star-merge. 
[06:23] <lalo> lulu: not sure yet, I'll tell you in a few minutes after I'm done merging :-P
[06:23] <lulu> lalo:good to hear you're merging! :o)
[06:25] <lalo> oh, found the thing I wanted in the merge. Then I can answer you, all is well :-D
[06:26] <lalo> I was worried I had introduced a bug Friday that would require me a few hours of hunting down - but no, it was just a typo and is already fixed by Daf
[06:28] <carlos> lalo: we are using now bugzilla to track the pending tasks/bugs
[06:29] <carlos> lalo: I suppose your mailbox knows it already  :-P
[06:29] <lalo> I know
[06:29] <lalo> yes :-)
[06:32] <daf> lalo: good to have you back :)
[06:33] <lalo> daf: thanks
[06:34] <lalo> but I wasn't gone for long, just an unexpectedly extended holiday :-P
[06:38] <lalo> hey Mark
[06:40] <daf> carlos: can you check whether #1912 is fixed?
[06:40] <carlos> daf: already done :-P
[06:41] <daf> carlos: and is it?
[06:41] <carlos> daf: the export still fails because the #1909
[06:41] <daf> oh, right
[06:41] <carlos> daf: yes, seems like it's working
[06:41] <daf> I'll change the server
[06:41] <daf> to use stub's branch
[06:42] <carlos> daf: which branch is it?
[06:42] <daf> stuart.bishop@canonical.com/zope--percentfix--3.0
[06:42] <carlos> ok
[06:42] <carlos> thanks
[06:42] <carlos> when will it be integrated into the default branch?
[06:42] <lalo> does the "make testdb" target still apply all necessary scripts to have a working testing database?
[06:43] <lalo> my ftests are failing with some funny postgres errors that suggest by db is incomplete
[06:43] <daf> carlos: when lifeless gets around to it
[06:43] <carlos> lalo: make launchpad_test at launchpad/database/schema should work
[06:44] <carlos> daf: ok
[06:44] <lalo> carlos: that's exactly what "make testdb" is a wrapper for :-/ doesn't work for me
[06:44] <lalo>     ProgrammingError: ERROR:  column translationeffortpotemplate.id does not exist
[06:45] <lalo> daf: yay, thanks!
[06:45] <daf> lalo: hmm?
[06:45] <lalo> your new bug explains the old one, I think :-)
[06:45] <daf> oh, right :)
[06:46] <lalo> I believe, when looking for whether to insert a new msgset or update an existing one, it must be missing the existing one due to it having sequence 0
[06:46] <carlos> lalo: do you have a sampledata-rosetta.sql inside the database directory and an empty sampledata.sql file ?
[06:46] <lalo> yes
[06:47] <carlos> then I don't know where is the problem, I did a rebuild of my database about 30 minutes ago...
[06:47] <carlos> it should work
[06:50] <daf> carlos: okay, the development server is now using stub's branch -- can you try again?
[06:50] <carlos> sure
[06:51] <carlos> daf: I get the file
[06:51] <daf> great!
[06:51] <daf> I'll close the bug
[06:51] <carlos> I need to check if it's correct :-)
[06:51] <carlos> daf: perfect
[06:52] <daf> :)
[06:52] <daf> carlos: can you also do some testing of submitting/updating translations through the web?
[06:52] <lalo> daf: do you have a preference on how to fix this? Should potemplate/pofile have methods that return a given msgset *even if* it has sequence 0?
[06:52] <carlos> hmm seems like all works execpt for this bug:
[06:52] <carlos> #: a11y/addressbook/ea-minicard-view.c:101
[06:52] <carlos> msgid "current addressbook folder"
[06:52] <carlos> msgstr[0]  "carpeta de libretas de direcciones actual"
[06:52] <lalo> this bug was, I as expected, introduced by removing the allowOld semantics :-)
[06:52] <carlos> all translations are exported that way 
[06:53] <lalo> carlos: oh
[06:53] <carlos> daf: yes, I will do it in some minutes
[06:53] <carlos> lalo: https://rosetta.warthogs.hbd.com/++skin++Debug/rosetta/projects/gnome/evolution/evolution-2.0/export.po
[06:53] <daf> lalo: well, since new translations from the web have sequence=0, you quite often need to deal with 0-sequenced message sets
[06:53] <lalo> carlos: I think the current way the adapters work is a bit baffling for the pofile's __unicode__ method
[06:53] <daf> carlos: if the export is broken, can you file a bug?
[06:53] <lalo> carlos: if there isn't a bug, please submit it, I'll fix next thing
[06:54] <carlos> daf: and the traversal algorithm I did fails O:-)
[06:54] <carlos> daf: sure
[06:54] <carlos> lalo: sure
[06:54] <carlos> :-)
[06:54] <carlos> daf: are you happy with the .po/.mo URL: https://rosetta.warthogs.hbd.com/++skin++Debug/rosetta/projects/gnome/evolution/evolution-2.0/es/po
[06:54] <lalo> daf: then pofile.__getitem__ should not require that sequence > 0
[06:54] <carlos> ?
[06:55] <lalo> what about potemplate? a template msgset with sequence 0 is, in theory, obsolete
[06:55] <daf> https://bugzilla.warthogs.hbd.com/bugzilla/showdependencygraph.cgi?id=1915 is fun
[06:55] <lalo> pofile.__getitem__ is ok (not broken), so this is just about templates
[06:55] <daf> carlos: yeah, that would be sufficient
[06:55] <carlos> ok
[06:56] <daf> lalo: I suspect that there are a few bogus "WHERE sequence > 0" in the SQL code
[06:56] <daf> I think I fixed a few
[06:58] <lalo> but in the case of potemplate.__getitem__ it's not bogus
[07:00] <lalo> well, before I do anything I must find why my database doesn't want to run the tests :-)
[07:02] <daf> lalo: how is your laptop now?
[07:02] <lalo> non-existent :-)
[07:02] <daf> lalo: how is your desktop now?
[07:02] <daf> :)
[07:03] <lalo> healthy
[07:03] <daf> that's good to hear
[07:03] <lalo> new mobo, 512 ram... there are still some some disagreements between hardware and drivers, but I'm not in a hurry to fix those
[07:05] <lalo> (eg my video card thinks my monitor is widescreen, so I'm at 1024x512 :-P but that doesn't affect work, so it's ok)
[07:18] <carlos> daf: I don't see a way to see the spaces when translating strings from rosetta...
[07:22] <daf> carlos: oh, yeah, I had forgotten about that
[07:23] <daf> carlos: I tried replacing all spaces with a space marker
[07:23] <daf> but that causes nasty problems with wrapping
[07:23] <daf> so I decided it would probably be ok to only mark leading and trailing spaces
[07:23] <lalo> well, I think I fixed it - now I have to become able to run the tests :-)
[07:23] <daf> but I haven't done that yet, so could you file a bug?
[07:24] <daf> lalo: fixed what?
[07:24] <carlos> daf: sure
[07:24] <lalo> aha
[07:24] <lalo> carlos: can you run the ftests?maybe my db is fine but the ftests are not
[07:25] <carlos> lalo: ok, just a minute I finish testing translations from rosetta
[07:25] <lalo> np
[07:26] <carlos> daf: ok, the translation works
[07:26] <carlos> daf: could you check it's already inside the database?
[07:26] <daf> carlos: hmm?
[07:27] <daf> carlos: if it's showing you the translations you put in, it works
[07:27] <daf> carlos: you could also test it by doing an export
[07:28] <carlos> daf: it's a way to test if it's not stored into the python objects
[07:28] <daf> carlos: sorry?
[07:29] <carlos> daf: I mean, I want to be sure that the information we get is stored into the database instead of being persistent data stored into python objects
[07:29] <carlos> that will be lost when launchpad is killed
[07:32] <daf> carlos: I think the database object are destroyed at the end of the request
[07:32] <daf> carlos: sorry, Python objects
[07:32] <carlos> ok
[07:32] <carlos> daf: I detected some bugs, filling them now
[07:32] <daf> great, thanks
[07:42] <lalo> pff.
[07:42] <lalo> daf: re "fixed what?" - 1905
[07:43] <daf> lalo: ah, great!
[07:44] <daf> what was the problem?
[07:45] <lalo> what I just said
[07:45] <lalo>         try:
[07:45] <lalo>             msgset = self.potemplate[msgid] 
[07:45] <lalo>         except KeyError:
[07:45] <lalo>             msgset = self.potemplate.createMessageSetFromText(msgid)
[07:45] <daf> oh, I see
[07:45] <lalo> potemplate.__getitem__ ignores sets with sequence=0, so it will insert a duplicate
[07:46] <lalo> I added a new method that doesn't :-( do minimize code duplication, I made __getitem__ a wrapper around this new method
[07:46] <daf> potemplate.createMessageSetFromText should have probably check against duplication
[07:47] <lalo> hmm yes
[07:47] <lalo> the code that did that was removed together with allowOld :-P I'll re-add it
[07:47] <lalo> ok, cya tomorrow
[07:47] <daf> I'll be back later
[07:47] <lalo> cya later then :-)
[07:47] <daf> and I'll probably be up until some strange hour in the morning :)
[07:48] <carlos> daf: later!!
[07:48] <lalo> actually the method shouldn't *have to* check for duplicates... the db should be raising an error
[07:49] <daf> lalo: hmm, if it's possible for the database to do that, it should
[07:49] <carlos> lalo: there are still some unique keys missing
[07:49] <lalo> daf: that's what unique is for
[07:50] <lalo> and createMessageIDSighting() should *definitely* "allowOld" by default - everything will become too hairy if it doesn't
[07:50] <SteveA> sabdf1: ping.  just tried to phone you.
[07:51] <lalo> or maybe not, dunno. We'll get back to it later if it causes a problem :-P (yagni)
[07:51] <daf> I'm getting concerned about update=False and allowOld=False and things like this
[07:52] <daf> but I'm not going to worry about it too much now
[07:52] <lalo> getting concerned? I thought these were already removed from existence
[07:52] <daf> I think there are still a few left
[07:53] <daf> I'm also a bit worried about having .encode('utf-8') sprinkled over the code
[07:53] <lalo> we shouldn't
[07:53] <daf> carlos: #1938 and #1939 are interesting
[07:53] <lalo> technically sqlo is already fixed wrt unicode
[07:53] <daf> no, we shouldn't -- that's why I'm worried :)
[07:54] <lalo> you can, if you wish, schedule a task to "remove these one by one and see which ones are still necessary"
[07:54] <carlos> phone
[07:54] <lalo> then find why they are necessary and report them to spiv :-)
[07:54] <lalo> I suspect we will find that they can all be gone
[07:55] <spiv> Hmm? :)
[07:55] <spiv> Where are these?
[07:55] <lalo> places where we avoid passing unicodes to sqlo
[07:55] <lalo> I think as of now sqlo is already smart enough to deal with unicodes all the time
[07:58] <lalo> yeah, well, 1905 is fixed indeed - the test still doesn't pass due to 1935 :-(
[08:24] <lalo> carlos: I need a way to test 1936 locally :-)
[08:25] <carlos> lalo: change rosetta... with localhost :-P
[08:25] <lalo> are the hooks in-place enough that that would work?
[08:26] <lalo> traversal hooks I mean
[08:26] <lalo> what a silly question, seeing as rosetta.w.h.c runs from rocketfuel :-P
[08:27] <carlos> lalo: the export.po is not the final traversal we will use
[08:27] <carlos> but it works for testing the export code
[08:27] <carlos> it will be fixed always to the Spanish translation
[08:27] <lalo> ok
[08:28] <lalo> thx
[08:28] <lalo> anyway I'll first prod the ftest till it runs
[08:29] <lalo> fortunately the ftest output is smart enough to tell me whether the bug is fixed :-)
[08:30] <carlos> I'm going to test it now here (I was changing somethings from my launchpad installation)
[08:30] <lalo> the ftests are lost wrt. RosettaProject vs. xxxRosettaProject
[08:31] <carlos> lalo: FAILED (failures=3, errors=2)
[08:31] <carlos> lalo: I think you should use DBProject instead of xxxRosettaProject
[08:31] <carlos> that was part of the SteveA's move to global objects across launchpad applications
[08:31] <lalo> for the tests, whatever works is ok
[08:33] <SteveA> don't use the classes with xxx in the names
[08:33] <SteveA> they will be removed soon
[08:33] <SteveA> pretend they aren't there
[08:38] <carlos> SteveA: are you going to do it with the other objects or should we schedule a task to do it?
[08:38] <SteveA> I have many other classes done in the code on my machine
[08:39] <carlos> ok
[08:40] <SteveA> if it is blockingn any of your work, then file a bugzilla bug on it
[08:40] <lalo> carlos: 404?  o.O
[08:41] <carlos> lalo: :-?
[08:41] <lalo> ah, nm
[08:42] <lalo> I had an error, then inserted  the debug skin to see what the error was, and got 404
[08:42] <lalo> but it was because I typed "debug" wrong :-P
[08:42] <lalo> actually... what is the incantation to get debugging, again? Wasn't it either ++skin++debug or ++skin++Debug?
[08:43] <carlos> the later
[08:44] <carlos>  /++skin++Debug/rosetta....
[08:48] <lalo> 404 :-/ seems I have no skin named Debug
[08:50] <carlos> lalo: you need to copy a file
[08:50] <carlos> let me check it...
[08:51] <carlos> I think it's canonical.debugskin-configure.zcml
[08:52] <carlos> but I don't remember if it's from launchpad/package-includes to launchpad/lib/canonical/ or from the second to the first
[08:52] <carlos> lalo: check where do you have it and copy/link it to the other directory :-)
[08:52] <lalo> yes :-)
[08:53] <lalo> thanks
[08:54] <carlos> no problem 
[09:05] <lalo> I'm getting 1909 now
[09:06] <lalo> what the heck, the ftest is proof enough that the bug is fixed for me
[09:09] <carlos> lalo: it's fixed if you get zope from stuart's branch
[09:10] <lalo> ok, thanks
[09:24] <lalo> confirmed, 1936 fixed
[09:26] <carlos> lalo: perfect!!
[09:50] <lalo> ...and merged
[09:53] <carlos> lalo: if you detect any missing task, just add it there, please
[09:53] <carlos> blocking the 1915 if it should be fixed before the alpha release
[09:55] <lalo> ok. But I haven't, yet.
[09:55] <lalo> (detected missing tasks, I mean)
[09:56] <lalo> I think I'll try to unearth 1935
[10:20] <jblack> I'm not convinced that is even english.
[10:22] <carlos> jblack: X-)
[10:22] <jblack> The first half of it is fine, but then it dives off into lawyereese.
[10:40] <lalo> one down, one last to go :-P